Research map / Label-efficient and robust learning / Test-time adaptation
TIPI: Test Time Adaptation With Transformation Invariance
A. Tuan Nguyen, Thanh Nguyen-Tang, Ser-Nam Lim, Philip H.S. Torr
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use transformation-invariance adaptation instead of entropy minimization | Test-time adaptation on CIFAR10-C · CIFAR10-C (average) | accuracy: 65.5 -> 80.4 | Improved |
| Use transformation-invariance adaptation instead of entropy minimization | Test-time adaptation on CIFAR100-C · CIFAR100-C (average) | accuracy: 39.8 -> 63.2 | Improved |
| Use transformation-invariance adaptation instead of test-time batch statistics | Test-time adaptation on ImageNet-C · ImageNet-C (average) | accuracy: 47.2 -> 55.0 | Improved |
| Use transformation-invariance adaptation instead of entropy minimization | Test-time adaptation on CIFAR10-C · CIFAR10-C (batch size 2) | accuracy: 17.4 -> 78.6 | Improved |
| Use transformation-invariance adaptation instead of entropy minimization | Test-time adaptation on CIFAR100-C · CIFAR100-C (batch size 2) | accuracy: 1.5 -> 56.3 | Improved |
| Use transformation-invariance adaptation instead of entropy minimization | Test-time adaptation on ImageNet-C · ImageNet-C (batch size 2) | accuracy: 0.8 -> 45.2 | Improved |
| Use transformation-invariance adaptation instead of entropy minimization | SVHN-to-MNIST adaptation · SVHN→MNIST (batch size 128) | accuracy: 90.7 -> 91.6 | Improved |
| Use transformation-invariance adaptation instead of T3A | SVHN-to-MNIST-M adaptation · SVHN→MNIST-M (batch size 4) | accuracy: 60.7 -> 64.2 | Improved |
| Use transformation-invariance adaptation instead of entropy minimization | SVHN-to-USPS adaptation · SVHN→USPS (batch size 128) | accuracy: 90.4 -> 93.9 | Improved |
| Use transformation-invariance adaptation instead of test-time batch statistics | Test-time adaptation on VisDA17 · VisDA17 (average) | accuracy: 53.9 -> 61.7 | Improved |
| Use transformation-invariance adaptation instead of entropy minimization | Test-time adaptation on VisDA17 · VisDA17 (batch size 2) | accuracy: 12.3 -> 51.2 | Improved |
| Use reverse KL rather than forward KL in the invariance objective (ablation) | TIPI · Experiments across domains and tasks | accuracy | Improved |
Later papers that compared with this method
2 later accepted papers report TIPI as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- FRET: Feature Redundancy Elimination for Test Time Adaptation (ICCV 2025 · better on 3 of 3 shared results)
- Synchronizing Task Behavior: Aligning Multiple Tasks during Test-Time Training (ICCV 2025 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TENT — Tent: Fully Test-Time Adaptation by Entropy Minimization (ICLR 2021) · better on 9 of 9 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- A Layer Selection Approach to Test Time Adaptation (AAAI 2025 · similarity 0.87)
- Towards Real-World Test-Time Adaptation: Tri-net Self-Training with Balanced Normalization (AAAI 2024 · similarity 0.86)
- Unified Entropy Optimization for Open-Set Test-Time Adaptation (CVPR 2024 · similarity 0.86)
- Bayesian Weight Enhancement with Steady-State Adaptation for Test-time Adaptation in Dynamic Environments (ICML 2025 · similarity 0.86)
- COME: Test-time Adaption by Conservatively Minimizing Entropy (ICLR 2025 · similarity 0.86)
- Test-Time Model Adaptation with Only Forward Passes (ICML 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
